When using AI coding assistants to refactor symbols across large codebases (6k+ files), developers face a binary choice: precision (LSP-based tools) or efficiency (grep/ripgrep). Shebe attempts to address this trade-off by way of a good old BM25 index, which is surprisingly fast and efficient.
Comments
When using AI coding assistants to refactor symbols across large codebases (6k+ files), developers face a binary choice: precision (LSP-based tools) or efficiency (grep/ripgrep). Shebe attempts to address this trade-off by way of a good old BM25 index, which is surprisingly fast and efficient.